Gradient filled plot panels with facet

Question

I did the following barchart by using:

ggplot(yearplot,aes(x=reorder(vE,release.year), y=difference,fill=PAR,label=signif.))  + 
  geom_bar(stat='identity',position = "dodge",  size=3 ,alpha=1)  +
  geom_text(aes(x = vE, y = difference+0.03 , label = format(signif., nsmall = 0, scientific = FALSE)), 
  color="black",size=6) +
  xlab("Genotypen")+
  ylab("Differenz")+
  theme_bw()+
  ylim(-0.2,0.2)+
  facet_wrap(~number,labeller = labeller(number=date.wrap),nrow=1)+
  coord_flip()+
  scale_fill_gradient2(name="PAR \n[mol/m² s]",low="darkblue",mid = "orange1",high = "orangered4",limits=c(1200, 2100),midpoint = 1200)+
  theme(legend.position="top")+
  theme(legend.key.size = unit(3,"line"),
        legend.title = element_text(size = 16,face="bold"),
        legend.text = element_text(size = 14))+
  theme(axis.text=element_text(size=14,face="bold"),
        axis.title=element_text(size=16,face="bold"),
        strip.text = element_text(size=16,face="bold"))

I would like to add temperature values as gradient filled plot panels, similar to the PAR values. So for example the plot panel of june 1 is light green and june 6 is dark green and the rest between. In the data frame is one temperature value for each day. Is there any possibilty to do that?

Please provide a reproducible example of your data (see: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). Otherwise, it will be really complicated to help you address this problem — dc37, Jan 02 '20 at 22:37

score 0 · Accepted Answer · answered Jan 03 '20 at 01:24

Your question contains a lot of code, but it is nonetheless not reproducible. A shame! It's in general an interesting question. Although I personally think that 'less is more' and one should try not to put too much information in one's graphs, here is one solution.

You want plot panels filled? Make large rectangles underneath your bars.
You want two fills? Use ggnewscale.

Voilà!

library(tidyverse)
library(ggnewscale)

mtc <- tibble::rownames_to_column(mtcars, 'model')

rect_df <- data.frame(xmin = -Inf, xmax = +Inf, ymin = -Inf, ymax = +Inf, cyl = unique(mtc$cyl))

ggplot() +
  geom_rect(data = rect_df, aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax, fill = cyl)) +
  facet_grid(~cyl) +
  scale_fill_gradient(low="lightgreen",high = "darkgreen") +
  ggnewscale::new_scale_fill() +
  geom_col(data = mtc, aes(x = model, y = gear, fill = disp)) +
  scale_fill_gradient2(low="darkblue",mid = "orange1",high = "orangered4",limits=c(min(mtc$disp),max(mtc$disp)),midpoint = mean(mtc$disp)) +
  coord_flip()

^{Created on 2020-01-03 by the reprex package (v0.3.0)}

Gradient filled plot panels with facet

1 Answers1